Update app.py

app.py

@@ -1,104 +1,104 @@
-import streamlit as st
-import datetime
-import pandas as pd
-from gnews import GNews
-from transformers import pipeline
-import plotly.graph_objects as go
-
-# Load the sentiment analysis model
-pipe = pipeline("text-classification", model="pramudyalyza/bert-indonesian-finetuned-news-v2")
-
-# Function to process the keyword and get sentiment analysis
-def process_keyword(keyword):
-    one_week_ago = datetime.datetime.now() - datetime.timedelta(days=7)
-    
-    news = GNews(language='id', country='ID', max_results=100)
-    
-    search_results = news.get_news(keyword)
-    
-    filtered_headlines = []
-    for article in search_results:
-        published_date = datetime.datetime.strptime(article['published date'], '%a, %d %b %Y %H:%M:%S %Z')
-        if published_date > one_week_ago:
-            filtered_headlines.append(article['title'])
-    
-    df = pd.DataFrame(filtered_headlines, columns=['title'])
-    df_clean = df.drop_duplicates()
-
-    df_clean['sentiment'] = df_clean['title'].apply(lambda x: pipe(x)[0]['label'])
-
-    positive_count = (df_clean['sentiment'] == 'Positive').sum()
-    negative_count = (df_clean['sentiment'] == 'Negative').sum()
-    neutral_count = (df_clean['sentiment'] == 'Neutral').sum()
-    total_count = len(df_clean)
-
-    return positive_count, negative_count, neutral_count, total_count, df_clean
-
-# Streamlit app layout
-st.title("News Sentiment Analysis Dashboard")
-
-keyword_input = st.text_input("Enter a keyword to search for news", placeholder="Type a keyword...")
-
-if st.button("Analyze"):
-    if keyword_input:
-        with st.spinner('Scraping and analyzing the data...'):
-            positive_count, negative_count, neutral_count, total_count, df_clean = process_keyword(keyword_input)
-
-        # Create plots
-        fig_positive = go.Figure(go.Indicator(
-            mode="gauge+number",
-            value=positive_count,
-            title={'text': "Positive Sentiment"},
-            gauge={'axis': {'range': [0, total_count]},
-                   'bar': {'color': "green"}}
-        ))
-
-        fig_negative = go.Figure(go.Indicator(
-            mode="gauge+number",
-            value=negative_count,
-            title={'text': "Negative Sentiment"},
-            gauge={'axis': {'range': [0, total_count]},
-                   'bar': {'color': "red"}}
-        ))
-
-        fig_neutral = go.Figure(go.Indicator(
-            mode="gauge+number",
-            value=neutral_count,
-            title={'text': "Neutral Sentiment"},
-            gauge={'axis': {'range': [0, total_count]},
-                   'bar': {'color': "yellow"}}
-        ))
-
-        fig_donut = go.Figure(go.Pie(
-            labels=['Positive', 'Negative', 'Neutral'],
-            values=[positive_count, negative_count, neutral_count],
-            hole=0.5,
-            marker=dict(colors=['green', 'red', 'yellow'])
-        ))
-        fig_donut.update_layout(title_text='Sentiment Distribution')
-
-        # Create a horizontal layout using st.columns
-        col1, col2, col3 = st.columns(3)
-
-        # Display results in each column
-        col1.plotly_chart(fig_positive, use_container_width=True)
-        col2.plotly_chart(fig_negative, use_container_width=True)
-        col3.plotly_chart(fig_neutral, use_container_width=True)
-
-        st.plotly_chart(fig_donut, use_container_width=True)
-
-        st.write(f"News articles found: {total_count}")
-
-        # Show DataFrame
-        st.dataframe(df_clean, use_container_width=True)
-
-        # Download CSV
-        csv = df_clean.to_csv(index=False).encode('utf-8')
-        st.download_button(
-            label="Download CSV",
-            data=csv,
-            file_name='news_sentiment_analysis.csv',
-            mime='text/csv',
-        )
-    else:
+import streamlit as st
+import datetime
+import pandas as pd
+from gnews import GNews
+from transformers import pipeline
+import plotly.graph_objects as go
+
+# Load the sentiment analysis model
+pipe = pipeline("text-classification", model="pramudyalyza/bert-indonesian-finetuned-news")
+
+# Function to process the keyword and get sentiment analysis
+def process_keyword(keyword):
+    one_week_ago = datetime.datetime.now() - datetime.timedelta(days=7)
+    
+    news = GNews(language='id', country='ID', max_results=100)
+    
+    search_results = news.get_news(keyword)
+    
+    filtered_headlines = []
+    for article in search_results:
+        published_date = datetime.datetime.strptime(article['published date'], '%a, %d %b %Y %H:%M:%S %Z')
+        if published_date > one_week_ago:
+            filtered_headlines.append(article['title'])
+    
+    df = pd.DataFrame(filtered_headlines, columns=['title'])
+    df_clean = df.drop_duplicates()
+
+    df_clean['sentiment'] = df_clean['title'].apply(lambda x: pipe(x)[0]['label'])
+
+    positive_count = (df_clean['sentiment'] == 'Positive').sum()
+    negative_count = (df_clean['sentiment'] == 'Negative').sum()
+    neutral_count = (df_clean['sentiment'] == 'Neutral').sum()
+    total_count = len(df_clean)
+
+    return positive_count, negative_count, neutral_count, total_count, df_clean
+
+# Streamlit app layout
+st.title("News Sentiment Analysis Dashboard")
+
+keyword_input = st.text_input("Enter a keyword to search for news", placeholder="Type a keyword...")
+
+if st.button("Analyze"):
+    if keyword_input:
+        with st.spinner('Scraping and analyzing the data...'):
+            positive_count, negative_count, neutral_count, total_count, df_clean = process_keyword(keyword_input)
+
+        # Create plots
+        fig_positive = go.Figure(go.Indicator(
+            mode="gauge+number",
+            value=positive_count,
+            title={'text': "Positive Sentiment"},
+            gauge={'axis': {'range': [0, total_count]},
+                   'bar': {'color': "green"}}
+        ))
+
+        fig_negative = go.Figure(go.Indicator(
+            mode="gauge+number",
+            value=negative_count,
+            title={'text': "Negative Sentiment"},
+            gauge={'axis': {'range': [0, total_count]},
+                   'bar': {'color': "red"}}
+        ))
+
+        fig_neutral = go.Figure(go.Indicator(
+            mode="gauge+number",
+            value=neutral_count,
+            title={'text': "Neutral Sentiment"},
+            gauge={'axis': {'range': [0, total_count]},
+                   'bar': {'color': "yellow"}}
+        ))
+
+        fig_donut = go.Figure(go.Pie(
+            labels=['Positive', 'Negative', 'Neutral'],
+            values=[positive_count, negative_count, neutral_count],
+            hole=0.5,
+            marker=dict(colors=['green', 'red', 'yellow'])
+        ))
+        fig_donut.update_layout(title_text='Sentiment Distribution')
+
+        # Create a horizontal layout using st.columns
+        col1, col2, col3 = st.columns(3)
+
+        # Display results in each column
+        col1.plotly_chart(fig_positive, use_container_width=True)
+        col2.plotly_chart(fig_negative, use_container_width=True)
+        col3.plotly_chart(fig_neutral, use_container_width=True)
+
+        st.plotly_chart(fig_donut, use_container_width=True)
+
+        st.write(f"News articles found: {total_count}")
+
+        # Show DataFrame
+        st.dataframe(df_clean, use_container_width=True)
+
+        # Download CSV
+        csv = df_clean.to_csv(index=False).encode('utf-8')
+        st.download_button(
+            label="Download CSV",
+            data=csv,
+            file_name='news_sentiment_analysis.csv',
+            mime='text/csv',
+        )
+    else:
         st.error("Please enter a keyword.")